Assigning cells to switches in cellular mobile networks: a comparative study

Assigning cells to switches in cellular mobile networks is an NP-hard problem which, for realsize mobile networks, could not be solved by using exact methods. In this context, heuristic approaches like genetic algorithms, tabu search and simulated annealing can be used. This paper proposes a comparative study of three heuristics—tabu search, simulated annealing and Parallel Genetic Algorithms with Migrations (PGAM)—used to solve this problem. The implementation of these algorithms have been tested in order to measure the quality of solutions. The results obtained confirm the efficiency of the seheuristics to provide good solutions for mediumand large-sized cellular mobile networks, in comparison with the sequential genetic algorithm (SGA) and with other heuristic methods well known in the literature. The evaluation cost provided by tabu search, simulated annealing and PGAM are very similar. Those three algorithms provide better results than standard genetic algorithm and always find feasible solutions. In terms of CPU times, tabu search is the fastest method. Finally, the results have been compared with a lower bound on the global optimum; they confirm the effectiveness of the tabu search, simulated annealing and PGAM to solve this problem. q 2003 Elsevier Science B.V. All rights reserved.